bertram 1

8/8/2019 Bertram 1

1/23

Institutionalising the Difference Principle

Christopher Bertram

(Department of Philosophy, University of Bristol)

(Paper for the Nuffield Political Group, Monday, November

30, 1998)1

INTRODUCTION

In a recent paper, Andrew Williams has produced an ingenious

critique of Jerry Cohens attempt to undermine the incentives

argument for inequality.2 I, in turn, want to rebut his charges. I

shall argue that a principle that persons should exhibit concern

for the least advantaged in their market behaviour is more easily

institutionalisable that Williams allows. Then I go on to argue a

different point, namely that the distinction between structure

and ethos is one that should be abandoned when we consider which

form of society best operationalises the difference principle,

because of the way in which the functioning of structures depends

on ethos. But first a bit of scene-setting.

THE DIFFERENCE PRINCIPLE

The so-called difference principle, the subordinate second

part of Rawlss subordinate second principle of justice, has

8/8/2019 Bertram 1

2/23

C. Bertram/Institutionalising the Difference Principle page 2

perhaps generated as much criticism and commentary as the rest of

his oeuvre put together. The principle initially of multiply

ambiguous formulation enjoins that society be so-organised

(subject to the meeting of a number of prior conditions) that the

expectations of the least advantaged (whoever they turn out to

be) are maximised. Those expectations are, in turn, to be

measured in terms of a quantity of all-purpose means, called

primary goods, available to those least advantaged persons over

the course of a lifetime. That share, in the hands of those least

advantaged persons, is not, according to Rawls, to be distributed

directly, as an aim of public policy or at any rate, not in the

first instance. Rather, the economic distribution that emerges in

society is to emerge as the result of the day-to-day operations

of its basic structure, with tax and transfer policies being a

mere mopping up operation to secure what improvement can be

secured over that naturally emergent distribution. The difference

principle does not enjoin equality, but rather sanctions an

unequal distribution of primary goods. The thought behind this is

that the least advantaged will themselves benefit from an unequal

distribution because the productivity-enhancing effects of

economic incentives will be such as to increase the size of the

economic cake and, crucially, of their slice of it.

In this paper, I say nothing very much by way of justification

of the principle, since arguments on that score have been gone

over may times. My interest in the body of the paper is to

intervene in a debate that goes to the heart of Rawlss theory,

namely G.A. Cohens challenge to the idea that for distributive

justice at least Rawlss is wrong to restrict his focus to the

basic structure as the primary subject of justice and should

8/8/2019 Bertram 1

3/23


instead take account of the behaviour and motivations of persons

within a well-ordered society and of the distributive patterns

that depend on their behaviour and motivations. I shall argue,

with Cohen and contra some of his critics such as Andrew

Williams that Cohens expansion of concern beyond the basic

structure of society is amply justified. But I shall also argue

that there is more to be said on the subject of

institutionalisation of an egalitarian ethos than Williams

allows. I then go on to argue against the very possibility of

identifying a difference-principle optimal basic structure in

ignorance of the motivations of the persons who are to operate

within in. The case of blood-transfusion familiar from the work

of Titmuss will provide a familiar example.3

THE BASIC STRUCTURE

As Brian Barry and others have recognised,4 Rawlss A Theory of

Justice represented a massive leap forward in the discourse of

liberal political theory in one significant respect (among many

others) in that it fully took on board the lesson of Marx and of

the classical sociologists, that social structure is profoundly

fateful for how peoples lives go. People, to paraphrase Marx,

may well make their own lives, but not in circumstances of their

own choosing, but rather within social and political structures

that they have to take as pretty much given at least in normal

circumstances. Rawls focused on that structure and asked what its

character would be if it were to be freely chosen in the light of

our nature as moral beings, rather than imposed upon us. The

whole thought-experiment of the original position is aimed at

8/8/2019 Bertram 1

4/23


providing design principles for such a structure. In focusing on

the structure of society in this way, Rawls achieved a remarkable

ideological victory on two fronts: first he responded to the

challenge of the left which had long correctly recognised the

importance of structure and, second, he blunted the challenge

of classical liberalism and libertarianism. He blunted this

challenge by marrying justice in process and justice in outcome

together through a legal, political and economic structure so

that he was less vulnerable to the challenge mounted by Hayek

(and later by Nozick) that social-democratic redistribution

required constant and unjustifiable interference with everyday

human activity. Rather, he sought both to guarantee essential

freedoms and to construct permissible processes in such a way as

to lead to desired distributive outcomes.

One might remark at this stage that Rawls appears to have a

great deal of confidence as was perhaps appropriate during the

post-war boom in the ability of economic experts to work out

which structure is to the long-term advantage of the least

advantaged. Given a lower degree of confidence in that ability

as is certainly appropriate today it may be rational to favour

more egalitarian short-term solutions because of a heavy rate of

discount of long-term egalitarian predictions.

COHENS CHALLENGE

When Rawls invites us to think about the problem of social

justice, he invites us to consider three perspectives: first, our

own considered judgements about the matter in hand; second, his

own contractarian construction the original position which is

8/8/2019 Bertram 1

5/23


supposed to be a procedural embodiment of those judgements; and

third, the viewpoint of the well-ordered society. The well-

ordered society is an imagined institutional realisation of

principles of justice where citizens act from a commitment to the

principles of justice.

Now Cohens key point is this. Given that such citizens have a

commitment to the principles of justice, how could it be the case

that they require productivity-enhancing incentives to supply

their skills in the amounts necessary to be of the greatest

benefit to the least advantaged? To be sure, some inequalities of

income might be needed: such as those strictly causally necessary

to elicit the required performance and those needed to compensate

people for the intrinsic unpleasantness of certain types of work.

But the latter, at least, is perhaps best understood anyway as

the fulfilment of what equality requires rather than a deviation

from it. What Cohen is above all concerned to rule out, as

incompatible with a commitment to justice, is the attitude of the

hard-bargaining high-flier in the market, who says, in effect, I

will only supply my scarce skills, if you pay me n, where n is

some figure far in excess of what the person actually needs in

order to perform the work in question. When a person takes such

an attitude, it may well be prudent to accede to their demands,

but their making those demands is revealing of a lack of

fraternity with the putative beneficiaries of the supply of those

skills the least advantaged.

One reply to Cohens gambit is to insist on its irrelevance to

the critique of Rawls, on the grounds that Rawlss focus is the

basic structure of society and not the attitudes of its members,

so long as those members conform to the requirements of that

8/8/2019 Bertram 1

6/23


structure. Such an approach runs up against Rawlss numerous

obiter dicta about the motivation of citizens, fraternity, and so

on. But as Cohen has pointed out, it also presents Rawlsians with

a dilemma. The first horn of that dilemma is that if one

conceives of the basic structure of society as simply its formal

legal structure, then the defence of Rawls is straightforward but

costly. Rawls could indeed maintain the position that incentive-

generated inequality is consistent with justice understood as

legal conformity, but given the fatefulness for human lives of

social phenomena more extensive than the basic structure, the

restriction of focus would seem arbitrary and indeed subversive

of the point of a theory of social justice. The second horn is

that if one moves beyond formal legal structure to the wider

network of social practices (as the requirement of fatefulness

would seen to enjoin) then it because much more difficult to

distinguish structure from behaviour since structures (e.g. of

the family) are often constituted by behavioural regularities.

The patriarchal family has the structure that it has because of

the behaviours of the members that constitute it.

The great strength of Andrew Williamss critique of Cohens

critique of Rawls is that, taking as its starting point the same

vantage point as Cohen namely, the perspective of the well-

ordered society Williams manages an elaboration of the basic

structure objection which permits both a more extensive construal

of the basic structure than the merely legalistic and a

restriction of its focus to exclude the motivation of law-abiding

agents in the marketplace. Williams does this by an ingenious

combination of two factors. He refuses to accept a definition of

the basic structure in terms which are either purely

8/8/2019 Bertram 1

7/23


dispositional (that which is fateful for peoples lives) or in

terms of a list of intrinsic properties (coercive legal

relations). Instead he opts for a mixed definition, claiming that

Rawlss account of social unity in a well-ordered society gives

us reason to restrict the scope of the basic structure to those

aspects of society that are both (a) fateful and (b)

institutionalisable in public systems of rules. The key point

here is that in a well-ordered society, justice must not only be

done but must be seen to be done. Public systems of norms hold

open the possibility for citizens of mutual verification that

their actions are in conformity with the shared social framework

that they are committed to and which they take to be expressive

of their moral personality. By contrast, the motives of agents in

the marketplace are often obscure (even to the agents themselves)

and so fall outside the purview of justice. In other words,

Rawlss cut between elements which fall within and without the

basic structure is not an arbitrary one, but one mandated by

other important elements within his theory.

PUBLICITY AND INSTITUTIONALIZATION

Andrew Williams has made the publicity requirement central to

his critique of Cohen because of the role that the requirement

plays in the possibility of well-ordered social co-operation.

According to that ideal, Williams explains,a society is well-

ordered only if regulated by a conception of justice that is both

public and stable. Under such conditions everyone accepts and

knows that others accept the same conception, and everyone knows

that conception is satisfied. Furthermore, everyone willingly

8/8/2019 Bertram 1

8/23


complied with the conception because, having witnessed others

readiness to act justly, they have internalized its requirements,

which in turn are congruent with their other values.5 According

to Williams, then, the institutional rules that comprise the

basic structure must be public in three respects: individuals

must be able to attain common knowledge of the rules (i)

general applicability, (ii) their particular requirements, and

(iii) the extent to which individuals conform with those

requirements.6 Williams regards as clearly disqualified both

self-effacing moral principles and those norms which are so

informationally demanding that individuals are incapable of

mutually verifying the status of their conduct.7

Now it can be difficult to reason about the character of an

ideally-just or well-ordered society. Such reasoning at the same

time neither purely a priori nor just a matter of empirical

generalisation. It isnt for example, part of the very idea of an

ideally just society that all the various desiderata of justice

are completely satisfied, since given conflicts and trade-offs

among those desiderata the maximally just society may

nevertheless remain defective when considered purely from the

perspective of liberty, equality or publicity. And rather like

those other desiderata, publicity is something which admits of

degrees of satisfaction. It is rather easier, for example, for

citizens to verify mutual compliance with some rules than with

others. In the case of highly complex rules, like those governing

corporate fraud, there may be some significant variation in

citizens ability to grasp either the rules general

applicability or their particular requirements. Given that many

rules playing an important function in securing the basic

8/8/2019 Bertram 1

9/23


liberties, guaranteeing equality of opportunity or maximinning

the advantage of the least advantaged will inevitably fail to be

maximally public, it would be heroic to insist on the maximal

publicity of rules as a necessary condition for a society to

count as just. This at least opens the door to norms, including

those derived from egalitarian ethi, which encounter some

difficulties on the road to full publicity.

It is worth noting that John Rawlss own remarks about the

institutional character of the basic structure are more nuanced

or at any rate more vague and evasive than Williams

interpretation might lead us to believe. Williams cites Rawlss

definition of an institution as a public system of rules and

reminds us that Rawls does not regard all norm-governed activity

as institutional. Instead he reserves the term for activity which

realises a certain type of norm, which is, in some sense,

public.8

But Rawls goes on to tell us that that which is

definitive of publicity here and most importantly the

possibility of mutual verification of compliance is not always

a feature of actual institutions but is an, albeit reasonable,

simplifying assumption. Now where institutions consist to some

degree of non-public rules, Rawls thinks that is to be deplored

(and other things being equal I need not disagree with him).

But it surely remains the case that those institutions whose

guiding norms are wholly or partly non-public still remain a part

of the basic structure of the wider societies that they inhabit.

How difficult is it going to be for a norm mandating that people

aim through their actions in the economic arena to bring about

the greatest benefit of the least advantaged to meet Williamss

publicity conditions? We can address this issue from a number of

8/8/2019 Bertram 1

10/23


directions. We could just ask how difficult it is in absolute

terms for such a norm to meet those conditions. But we can also

ask how this norm fares on the dimension of publicity compared to

other norms which Williams (and Rawls) want to see

institutionalised. If rules derived from an egalitarian ethos

fare no worse (or not much worse) on the publicity dimension than

other rules, then that will weigh against Williamss arguments.

And we can ask whether an egalitarian ethos which could not

itself satisfy those publicity requirements (to an adequate

degree) might nevertheless justify more concrete rules which

could meet those requirements.

Mutual verifiability

One of Williamss three criteria of publicity is the mutual

verifiability requirement. Now Im somewhat sceptical about

Williamss attribution to Rawls of this part of the publicity

requirement being a necessary condition for a norm to be embodied

in the basic structure. After all, if our model is that of a

well-ordered society, we already know by hypothesis that

individuals are motivated to act in accordance with principles of

justice, so the need to check whether they are in fact conforming

with the rules is pre-empted by an assumption of social trust.9

Finding out whether or not norms are being complied with is, in

any case, afflicted with difficulties of two different kinds, one

of which is much more important than the other. We can

distinguish between the difficulty of finding out whether or not

there are many acts which violate rule R and the difficulty of

knowing whether or not an act A is of a type which violates rule

R. So, for example, it might be very difficult to tell (for

8/8/2019 Bertram 1

11/23


obvious reasons) whether a prohibition on some sexual practice is

widely complied with, but it might be easy to tell of some

particular act (observed or indulged in) whether or not it is an

act of the putatively prohibited type. We can also distinguish

between first- and third-person perspectives on an act. Equal

opportunity legislation may leave little doubt about which acts

are permitted and which prohibited, but it may be hard for a

third-party to tell (and harder to establish) whether a given act

is an act violating the rule. A familiar example of this would be

when a pregnant woman is not appointed to a post. Denial of

employment on grounds of pregnancy would clearly be in breach of

the law, but the reasons publicly given by the employer are not

those which we suspect were actually operative.

We should in any case be wary of excessive scepticism concerning

the possibility of establishing whether a given other person is

conforming to a rule. For most offences in the criminal law by

anyones lights part of the core of the basic structure it is

necessary for the prosecution to establish not only that the

accused exhibited the behaviour specified by the law (the actus

reus) but also that they had the mental state appropriate for the

act in question (the mens rea). So, for example, a prosecution

for theft has to establish not only that a person, say, left a

shop with some unpaid-for goods, but also that they intended

permanently to deprive the rightful owner of their property.

Other laws explicitly build in reference to intention. There is,

for instance, a law against driving without due care and

attention. In all of these cases we draw inferences from public

behaviour and other facts concerning the mental state of the

putative offender there is not, in practice, a great difficulty

8/8/2019 Bertram 1

12/23


about institutionalising such laws and nor need there be about

institutionalising other rules.

Particular requirements

All of the above leads me not to dismiss entirely, but at least

to heavily downgrade the importance of mutual verifiability of

compliance with norms in a well-ordered society. But a much

greater challenge is posed by the need for citizens to know what,

specifically, the rules demand of them. Here again, there are

ambiguities. Is it enough, to satisfy this condition, that there

is for a rule a means of telling for each act whether or not it

is in conformity? Clearly not, for we can imagine rules for which

that condition is satisfied but only in ways that place

intolerable burdens on citizens reasoning or information-

handling capacities. (Perhaps, to establish whether on not a

given act is in conformity with the rule, citizens would have to

follow a long and complicated algorithm.) Do we want to say that

the condition is met if citizens can establish whether or not

their conduct complies with the rule so long as they consult an

expert (a lawyer, accountant, priest or sociologist)? If such

consultation is expensive, does that mean that the publicity

condition is met to any lesser degree?

There are surely many instances where citizens may find it

difficult or costly to discover whether or not their behaviour is

in conformity with a rule. But it may be possible for them to

find out something else, namely, whether their behaviour is in a

zone of risk. Citizens may not be sure whether or not a given

gesture is, technically speaking, an assault, or whether or not a

given piece of financial sharp practice counts, strictly

8/8/2019 Bertram 1

13/23


speaking, as fraud. But they do know that there is a general

principle fraud is wrong which finds expression in more or

less detailed regulations. The general principle does not itself

issue in specific rules, but specific rules are needed to give it

detailed expression. If we just consider the rules by themselves,

they dont seem to meet the publicity requirement, since they are

too informationally demanding. If we consider the principle

alone, it doesnt meet the conditions of publicity either because

it is too vague. But take the two together and we have an

ensemble which is capable of guiding conduct to a sufficiently

tolerable degree.

How does a norm mandating that people act so as to maximise the

advantage of the least advantaged fare? As Williams points out,

there are clear difficulties with meeting the condition that

citizens acquire knowledge of its particular requirements and

consequent problems that arise for mutual verifiability of

conformity. Some of these difficulties arise from the connection

between the principle and occupational choice: does the norm

require that I work in the job where my work will be to the

greatest benefit of the least advantaged? Other difficulties

arise from the egalitarian compensation requirement: am I due

extra pay to compensate me for the intrinsic unpleasantness of my

work? And some arise from the interaction between the principle

and reasonable agent-centred prerogatives.10 All of these

considerations suggest as Williams argues that Cohens

egalitarian ethos cannot be directly institutionalised in the

form of public rules that could form a part of the basic

structure. But that may not be to say anything very damaging: the

basic principles of justice themselves, and close derivatives of

8/8/2019 Bertram 1

14/23


them (such as the prohibition on fraud) are in a similar

position. The difficulties establish that institutionalisation is

not a straightforward matter and that there may be morally

required and epistemically inescapable limits to it, they dont

establish that nothing can be done.

What can be done? Whether or not there is a principled moral

distinction to be made between acts and omissions, it does seem

to be the case that the law (as one system of rules) finds acts

easier to deal with than failures to act: it is generally easier

to establish that A harmed B (when A did) than that A failed in

her duty to aid B. We certainly should not criminalise the person

who simply fails to assist the least advantaged to the greatest

possible degree. But that should not prevent us from establishing

a prohibition on knowingly conducting oneself in a way seriously

harmful to expectations of the least advantaged. If this norm

were embodied in law, convictions would no doubt be hard to

secure, especially given the need to prove intent. Since the case

would be difficult to prove, it would be hard for talented high-

fliers to complain that they were being unfairly placed in

jeopardy and, in any case, such fears would have to be balanced

against the long-term interests of the least advantaged. Even if

there were no cases ever, such a law would both fulfil an

important symbolic purpose: a society that implemented it would

signal its commitment to the principle of fraternity. (Exemplary

use of the criminal law is just one possibility, though. The

British Government of Tony Blair has already made use of

windfall taxation in order to seize the excessive profits of

privatised utility companies similar measures could perhaps be

employed to seize excessive personal income!)

8/8/2019 Bertram 1

15/23


A final suggestion for how we might implement an egalitarian

ethos concerns the behaviour of firms and other institutions. To

be sure, the ethos is an ethos governing the conduct of

individuals, but we may further it and establish its status as a

customary rule by placing requirement on employers. We could

legislate to require firms to have due regard to the satisfaction

of the difference principle in their employment practices and

obliging them to report on how they implemented that requirement.

Experience of similar laws concerning the environmental impact of

a companys policy has not been the catalogue of cynical evasion

that might have been expected. Of course, it might be difficult

to determine which employment practices would have the desired

effect. But given the reporting requirement we might hope to

build up a body of knowledge over time which could come to guide

decisions. Remuneration committees of large firms would have to

show how their decisions were consistent with the requirement and

we might expect overly cynical rationalisations to be the focus

of media scrutiny.

In all of these attempts at institutionalisation, what we have

to bear in mind is the weight that should be given to the various

desiderata of social justice. A basic structure which implements

an egalitarian ethos may, despite what I have said, do worse on

the dimension of publicity that one which does not. But the

principle of publicity is only instrumental to the goal of well-

ordered social co-operation and it may be more or equally

damaging to the achievement of that goal if the disadvantaged

come to resent the cynical and exploitative behaviour of high-

flying market maximisers.

8/8/2019 Bertram 1

16/23


ETHOS AND STRUCTURE

The so-called basic structure rebuttal of Cohen relies upon

the drawing of a sharp line between structure and ethos. Up till

now my aim has been mainly negative, I have been seeking to rebut

Williamss critique of Cohen. But I want to move on to raise a

further problem for those who seek to maintain a basic structure

ethos distinction. I want to point to one way in which structure

depends on ethos and then argue that the identification of the

difference-principle optimal structure, an identification which

the implementation of the difference principle requires, cannot

be done without taking account of the prevailing ethos. This is a

less radical approach than is involved in Cohens attempt to blur

the distinction between structure and ethos in two respects.

First, while for Cohen the problem is the deep one that some

kinds of structure are constituted by behavioural regularities,

my claim is that certain structures depend for their (effective)

functioning on ethi of certain kinds. Second, mine is more of a

rhetorical point against someone who believes that ethos is

irrelevant to justice. Someone who rather believes that people

have an obligation to act within the limits of psychological

and physical necessity and making due allowance for agent-centred

prerogatives in order to maximise the expectations of the least

advantaged, will not be concerned by what I have to say here.11

My thought is this: that the functioning of objectively

described basic structures will in many cases be a function of a

prevailing ethos. (I should add for the sake of completeness the

Rousseauean thought that the prevailing ethos may be the

consequence of the structure.) Let me say a little, by way of

elaboration. The basic structure that obtains in a Rawlsian just

8/8/2019 Bertram 1

17/23


society will emphatically not be a regime of free markets plus

tax and transfer. Rather, Rawls is clear that the ideal of a

property-owning democracy is to be instantiated through an

economic and social system which is likely to be a mix of

different forms of tenure, organisation, property law, and so on.

In any given property-owning democracy, we may expect to find

state-owned industries, firms operating employee share-ownership

schemes, worker co-operatives, private companies and so on. And

there will also be much room for variation in the extent to which

various goods and services are provided as commodities through

the market or as direct benefits from the state, or by people,

their social networks and families.12

Allowing for that

multidimensional variation in possible basic structures, we can

imagine a space containing all the different possible basic

structures, perhaps a space resembling Borgess library of Babel

which Daniel Dennett makes such effective expository use of in

his Darwins Dangerous Idea.13 Now consider the following

question: which of these structures is optimal for the

satisfaction of Rawlss difference principle? I submit that there

is no answer to that question in abstraction from what the

prevailing ethos of society is.

Someone might disagree with that. They might think that the

contribution of structure and ethos to difference-principle

satisfaction is additive: a particular structure always

contributes a constant amount to the satisfaction of the

difference principle and a particular ethos also contributes a

certain amount. Under that assumption, we might imagine all the

structures lined up in space with the most difference-principle

friendly on the left and the least friendly on the right and the

8/8/2019 Bertram 1

18/23


ordering of those structures from left to right would not vary

however we changed our motivational assumptions about the people

working in them (although the situation of the least advantaged

would vary for each structure as the ethos was changed). That

invariance assumption is deeply implausible. A simple

illustration should bring this out: that of blood transfusion.

Richard Titmusss famous study of The Gift Relationship deals

with the motivation of donors to the UK Blood Transfusion

service.14 Titmuss argued, not that donors were motivated by pure

altruism, but rather by a sense of generalised reciprocity: they

ought to give because, after all, blood transfusion might be

something that they too would need in the future. Purely self-

seeking market-maximisers would not contribute blood, of course,

because they would receive neither monetary reward nor an

improvement in their own situation were they themselves to need a

transfusion. It is easy to see that, given the objective of

getting enough blood to transfuse needy patients, the best

structure will differ depending upon the patterns of motivation

that exist in society. There is no optimal structure to be picked

out independently of those motivations and an attempt by an

economist to do so using the standard tools of their trade would

often pick out a structure that would be suboptimal. Given the

predominance of a self-seeking motivation, a system of voluntary

blood donation would be a disaster and would not generate the

desired result: rather a market based system of incentive

(payment for blood supplied) or disincentive (denial of

transfusion to non-donors) would be called for. But where persons

are motivated by a sense of mutual responsibility and dependency,

voluntary donation is superior.

8/8/2019 Bertram 1

19/23


Any attempt to treat structure and ethos separately is, in any

case, going to come to grief on the social and psychological fact

of the profound effect of structure on ethos, a fact noted and

emphasised by many writers and borne out by common experience.

Public sector institutions in the UK which have introduced

internal markets or devolved budgeting systems have also

experienced profound changes in the motivation and ethos of their

staff. Where once colleagues in one university department would

gladly do a bit of teaching for another one on the basis of rough

reciprocity or even just goodwill, the introduction of cost

centres means that what was once informal exchange is now

accounted and costed. As a result, much activity that once took

place now does not. No doubt there are many issues to consider

when an institution considers whether it should run on a

decentralised market basis, as a central dictatorship, as a

federation of co-operatives etc, and I dont want to suggest that

one model is necessarily better than another in all

circumstances. What I do want to do is to deplore the treatment

of these structures as neutral technologies for turning

preferences into outcomes in a way that neglects the way in which

structures transform preferences.

If we look at Rawls himself, we find, I believe an ambivalence

about the relationship between ethos and structure. Rawls is

often concerned with the way in which a form of society interacts

with the motivations of citizens and indeed emphasises the

importance of institutions that in practice foster a sense of

justice. He tells us, in another of his obiter dicta that In

designing and reforming social arrangements one must, of course,

examine the schemes and tactics it [the institution] allows and

8/8/2019 Bertram 1

20/23


the forms of behaviour which it tends to encourage.15

When he

talks about operationalising this idea, though, he immediately

slips into discussion of the Smithian invisible hand which

implies a conception of the state as providing systems of

incentives and disincentives for rational persons in order to

bring about some desired result. Sometimes forms of behaviour are

taken by Rawls as consequences of social structure, sometimes as

more or less given. I believe that it makes more sense to treat

the two together as interacting. The trouble is, that despite

more than two centuries of social theory we still have little

idea how to do this: although we can see that such interaction

does take place. But the difficulty of operationalising such an

idea is no real excuse for persons like Rawls who do accept

the plasticity of individual motivation to rely, in their design

of social institutions, on a body of theory which denies it.16

BY WAY OF CONCLUSION

I have tried to argue that whilst we should accept Andrew

Williamss point that the goal of well-ordered social co-

operation requires that principles of justice should be

implemented as far as possible through public systems of rules

there may be more room for the implementation of norms preventing

maximising behaviour by high-fliers than he allows. In any case,

meeting the publicity requirement is a tough test not just for

such norms but for many others, including those that Williams

himself must see as central to a just society. The goal of well-

ordered social co-operation requires not just publicity, but also

fraternity, and in order to secure it some trade-offs will be

needed. In any case, because of the well-documented interaction

between structure and ethos, the very identification of the

8/8/2019 Bertram 1

21/23


difference-principle optimal basic structure cannot be done

without a consideration of the ethi that it will permit and

foster. Whatever the difficulties in implementing an egalitarian

ethos, Cohen is right in his contention that a theory of social

justice cannot plausibly restrict itself to the evaluation of

abstract systems of rules but must also range over individual

behaviour and motivation.

1 Very much work in progress, so please dont cite without

permission. Thanks to Andrew Williams for discussion and comments.

2 Andrew Williams, Incentives, Inequality and Publicity, Philosophy

and Public Affairs, 27, no. 3 (1998), pp.226-248. For Cohen see,

Incentives, Inequality and Community, in G.B. Petersen ed., The

Tanner Lectures on Human Values. Vol. 13 (Salt Lake City: University

of Utah Press, 1992) pp. 262329; The Pareto Argument for

Inequality, Social Philosophy and Policy12 (1995), pp. 16085; and

Where the Action Is: On the Site of Distributive Justice,

Philosophy and Public Affairs 26, no. 1 (Winter 1997) pp. 330.

3 Richard Titmuss, The Gift Relationship (London: George Allen &

Unwin, 1970).

4 For remarks along these lines see, e.g., Brian Barry, Justice as

Impartiality(Oxford: Oxford University Press, 1995), p. 214.

5 Williams, p. 245.

6 Williams, p. 234.

7 Ibid., p. 235.

8 Ibid., p. 234.

9 As Andrew Williams has pointed out to me, this assumes that the

relevant laws dont have the form, perform act P in circumstance q

if you know that sufficient others will do x in q. I do indeed make

this assumption and I believe that Williams should too, since laws of

8/8/2019 Bertram 1

22/23


such a form and it would be unusual for laws to be so framed

would be likely to be more informationally demanding than he permits.

10 For an extensive exploration of the impact of agent-centred

prerogatives on Cohens critiques of Rawls see David Estlund,

Liberalism, Equality and Fraternity in Cohens Critique of Rawls,

Journal of Political Philosophy, vol. 6, no. 1 (March 1998),

11 I dont think that there is much mileage in the following

paradoxical thought, commonplace among proponents of Victorian

capitalism: those genuinely concerned about the least advantaged

should not seek to aid them but should instead promote and exemplify

an ethos of individual self-reliance. A variant of this would be to

argue, contra-Cohen, that inequalities are to the benefit of the

least advantage but not because of their productivity-enhancing

effect on those that have, but rather on those that have not.

12See Krouse, R. and M. McPherson, Capitalism, Property-Owning

Democracy, and the Welfare State, in A. Guttmann ed., Democracy

and the Welfare State.

13 Daniel C. Dennett, Darwins Dangerous Idea (Hamondsworth: Allen

Lane, 1995) ch, 5; Jorge Luis Borges, The Library of Babel, in

Labyrinths (New York, 1962).

14 Richard Titmuss, The Gift Relationship (London: George Allen &

Unwin, 1970). My discussion draws heavily on Alan Carling, What Do

Socialists Want?, unpublished ms.

15A Theory of Justice, p. 57.

16 Perhaps this is too strong, since economists often say that any

(consistent) motivational pattern is representable in theory. What

can be said, though, is that one of conditions of the welfare

theorems of neoclassical economics [is] that agents exhibit no

concern for the interests of those with whom they interact,

Christopher W. Morris, The Relation Between Self-Interest and

8/8/2019 Bertram 1

23/23


Justice in Contractarian Ethics, Social Philosophy and Policy, check

exact reference. If social institutions have the effect of fostering

such concern and thus render possible forms of interaction and co-

operation denied to non-tuistic agents, that fact will go unremarked

in the models used by standard economics.

bertram 1

Documents